Re: Yet another serialization format?

On 6/30/11 11:52 AM, Markus Lanthaler wrote:
> On 06/30/2011 4:50 AM Kingsley Idehen wrote:
>> You are saying, in plain English, with regards to basic computer
>> science:
>>
>> Disambiguating a Name from an Address doesn't matter when constructing
>> Linked Data Structures.
> Basically, yes, that's what I'm saying. But I was also saying that if you would like to disambiguate them, you should make it easier to talk about the more common case and that's in my opinion talking about the Thing.

I don't recall this sort of journey re. computer science fundamentals 
pre network enhanced hyperlinking via HTTP.
>
>> If that were true, I wouldn't even be able to send this email.
> Has nothing to do with the current discussion but would be interested to know why that's the case.

Email and every other application are driven by "data access by 
reference". That's how programs are written. That's intrinsic to 
computing (all forms, even pre. Silicon).

>   I find that example a bit amusing because I'm entering my e-mail address every day in dozens of applications to identify *myself* without appending a +me or any other fragment.

That's a different conversation. But if you think about this a little, 
the following should resonate:

1. An email address is a mailto: scheme URI
2. mailto: URIs can be made to resolve to representations of their 
Referents via protocols such as Webfinger and Foingerpoint
3. a resolvable mailto: scheme URI takes email sender identification 
beyond literal values (why SPAM is so trivial to generate and so 
difficult to kill) such that you can actually use logic to verify the 
identity of the mail sender via a digital certificate
4. generation of the digital certificate doesn't require CA network
5. verification of the digital certificate doesn't require CA network.

All of that happens because when Names and Addresses are distinct you 
can do this:

1. de-reference mailto: scheme URI
2. see if the mailto scheme URI is in a formal relation with a Public Key.

Based on the above, you can tweak existing protocols such as TLS/HTTPS 
and S/MIME.

And that leads you to WebID which is a great example of Linked Data 
utility and the fundamentals of "data access by reference" combined with 
trust logic.

The only thing that is different today is that URIs allows us to operate 
at InterWeb scales when exploiting Linked Data structures.

> I think we won't find a consensus on this because there's no real true or false IMHO. Since it's not really related to the problem we are addressing, neither to the proposal I made I would suggest to put this discussion aside and move on.
>
No, we debate ourselves to real life implementations of our world views. 
WebID is a practical example of a problem based on these fundamentals.


Links:

1. http://www.w3.org/wiki/WebID .
>


-- 

Regards,

Kingsley Idehen 
President&  CEO
OpenLink Software
Web: http://www.openlinksw.com
Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca: kidehen

Received on Thursday, 30 June 2011 11:53:03 UTC